Dr-Eng.

Imam Muthohar
Dr. Istiarto
Definisi REGRESI…
 Regresi adalah suatu metode analisis statistik yang digunakan
untuk melihat pengaruh antara 2 atau lebih variabel 
variabel bebas (independent variable) dan variabel bergantung
(dependent variable).
 Jenis-jenis regresi:
 Regresi SEDERHANA (linier sederhana dan non-linier sederhana)
 Regresi BERGANDA (linier berganda dan non-linier berganda)
 Analisis regresi digunakan pada hampir semua bidang
kehidupan: ekonomi dan keuangan, teknik, pertanian, industri
dan ketenagakerjaan, pemerintahan, ilmu lingkungan, dsb
 untuk mengetahui hubungan sistem (relation), pendugaan
(estimation), dan peramalan (forecasting)
 Independent variables are variables which having an affect on dependent variable.
 Two basic assumption are made concerning the form of the independent variables.
 The predictive variables are assumed to be independent of each other
 All variables are assumed to be random variables with a normal distribution
Define the accuracy of the predictive equation. Some of the most
common measures that are used in analysis or validity test procedure:
 linear correlation test between independent variables
(X) and dependent variable (Y)
 the coefficient of partial correlation (r)
 the coefficient of multiple determination (R2)
 statistical test (t-test or F-test)  check of normal
distribution, colinearity-independency level
The weakness: be aware!
 if the data are scattered, it is difficult for using linear
regression analysis (correlation coefficient very small)
 independent variables are not independent each other
 colinearity
the coefficient of partial correlation (r)
what is the meaning (-1 < r < +1)?
when it achieves the best condition?

the coefficient of multiple determination (R2)

what is the meaning (0 < r2 < +1)?
when it achieves the best condition?
CASE STUDIES :
Case 1
From the home interview survey, the information between number of cars and daily number of
trips was describe as shown in the table.

Determine the trip generation

model based on the
relationship between both of
them. Which is categorized
as dependent and
independent variable?
Case 2
Result of trip attraction model at certain zone.
Y = 123.2 + 0.89 X1 (R2 = 0.900)
Y = 40.1 + 0.14 X2 + 0.61 X3 + 0.25 X4 (R2=0.925)
Y = - 1.7 + 2.57 X1 – 1.78 X4 (R2 = 0.996)

where Y is the number of work trips that attracted to the

zone; X1 is total employee; X2 is employee in industrial
sector; X3 is employee in trading sector; X4 is employee in
service sector. Discuss this analysis result!
Case 3
An analyst came up with the following regression equations and a
simple correlation matrix for 20 zones, as given:

Y = 50.5 + 0.80 X1 (Se = 210, R2 = 0.95, t = 34)

Y = 308 + 0.79 X2 (Se = 884, R2 = 0.88, t = 29)
Y = 52.7 + 0.85 X2 + 1.75 X3
(Se = 205, R2 = 0.98, t = 60; 22)
Y = -105 + 1.38 X2 - 0.4 X3 + 0.1 X4
(Se = 155, R2 = 0.97, t = 3;2;0.5)

where Y = trips produced; X1 = total population; X2 = blue-collar population;

X3 = red-collar population; and X4 = school-going children
Case 3 (cont’d)
Table of partial correlation matrix
Y X1 X2 X3 X4
Y 1.00 0.95 0.85 0.42 0.23
X1 1.00 0.92 0.53 0.22
X2 1.00 0.35 0.09
X3 1.00 0.12
X4 1.00

Comment on the suitability of these equations for use in a

transportation study!
 Case 4:
Consider the following data of trip generation in the zones
and their socio-economic condition. Define the appropriate
trip generation model.

Bangkitan Kendaraan Penduduk PDRB Produksi Pertanian

Zona
(ribuan smp/thn) (ribuan) (milyar rupiah) (ribu ton/thn)
1 408 964 590 1
2 396 687 802 1.6
3 423 594 910 1.577
4 440 917 745 8
5 174 84 605 45
6 333 523 692 787
7 357 526 789 829
8 341 696 571 2
9 465 546 967 621
10 345 662 620 1547

Y X1 X2 X3
Case 4 : Problem solving
 Check the linear correlation test between independent
variables (X) and dependent variable (Y)
 Calculation of multiple linear regression
 Check the parameter value which has the lowest correlation
– retrial to do the same process
 Determine the best model
Case 4 : Problem solving
Check the linear correlation test between independent
variables (X) and dependent variable (Y)

2 2
Bangkitan Kendaraan Penduduk y x xy x y
Zona
(ribuan smp/thn) (ribuan) 1 408 964 393,312 929,296 166,464
1 408 964 2 396 687 272,052 471,969 156,816
2 396 687 3 423 594 251,262 352,836 178,929
3 423 594 4 440 917 403,480 840,889 193,600
4 440 917 5 174 84 14,616 7,056 30,276
5 174 84 6 333 523 174,159 273,529 110,889
6 333 523 7 357 526 187,782 276,676 127,449
7 357 526 8 341 696 237,336 484,416 116,281
8 341 696 9 465 546 253,890 298,116 216,225
9 465 546 10 345 662 228,390 438,244 119,025
10 345 662 Σ 3,682 6,199 2,416,279 4,373,027 1,415,954

1,338,072 1,787,292
r 0.75 0.748665479

Y X
Case 4 : Problem solving
b0 + b1 X1 + b2 X2 + b3 X3 = Y

N N N N
Nb0  b1  X 1i  b2  X 2i  b3  X 3i   Yi
i 1 i 1 i 1 i 1
N N N N N
b0  X 1i  b1  ( X 1i )  b2  ( X 1i . X 2i )  b3  ( X 1i . X 3i )   ( X 1i .Yi )
2

i 1 i 1 i 1 i 1 i 1

N N N N N
b0  X 2i  b1  ( X 1i . X 2i )  b2  ( X 2i )  b3  ( X 2i . X 3i )   ( X 2i .Yi )
2

i 1 i 1 i 1 i 1 i 1

N N N N N
b0  X 3i  b1  ( X 1i . X 3i )  b2  ( X 2i . X 3i )  b3  ( X 3i )   ( X 3i .Yi )
2

i 1 i 1 i 1 i 1 i 1
 Case 4 : Problem solving

2 2 2
Zona Y X1 X2 X3 X1 X2 X3 X1.X2 X1.X3 X2.X3 X1.Y X2.Y X3.Y
1 408 964 590 1 929,296 348,100 1 568,760 964 590 393,312 240,720 408
2 396 687 802 2 471,969 643,204 3 550,974 1,099 1,283 272,052 317,592 634
3 423 594 910 2 352,836 828,100 2 540,540 937 1,435 251,262 384,930 667
4 440 917 745 8 840,889 555,025 64 683,165 7,336 5,960 403,480 327,800 3,520
5 174 84 605 45 7,056 366,025 2,025 50,820 3,780 27,225 14,616 105,270 7,830
6 333 523 692 787 273,529 478,864 619,369 361,916 411,601 544,604 174,159 230,436 262,071
7 357 526 789 829 276,676 622,521 687,241 415,014 436,054 654,081 187,782 281,673 295,953
8 341 696 571 2 484,416 326,041 4 397,416 1,392 1,142 237,336 194,711 682
9 465 546 967 621 298,116 935,089 385,641 527,982 339,066 600,507 253,890 449,655 288,765
10 345 662 620 1,547 438,244 384,400 2,393,209 410,440 1,024,114 959,140 228,390 213,900 533,715
Σ 3,682 6,199 7,291 3,843 4,373,027 5,487,369 4,087,559 4,507,027 2,226,343 2,795,967 2,416,279 2,746,687 1,394,245

10 b0 6,199 b1 7,291 b2 3,843 b3 = 3,682

6,199 b0 4,373,027 b1 4,507,027 b2 2,226,343 b3 = 2,416,279
7,291 b0 4,507,027 b1 5,487,369 b2 2,795,967 b3 = 2,746,687
3,843 b0 2,226,343 b1 2,795,967 b2 4,087,559 b3 = 1,394,245

Y = -77,428 + 0,264 X1 + 0,382 X2 + 0,0087 X3